A device which can capture all the packets flowing through the communication network over a long time is required in preparation for dealing with communication failures in communication networks such as the Ethernet (registered trademark). To satisfy this requirement, the data captured should be recorded in a secondary storage at high rate without being dropped.
Conventional data writing methods using parallel secondary storages have been proposed as a technique of writing data in secondary storages at high rate (for example, see Patent Literature (PTL) 1). According to PTL 1, secondary storages are disposed in parallel, and data is written in one of the secondary storages which is under the lowest load. By such a configuration, data input/output to and from the parallel secondary storages is increased without an imbalance in access load.